Herbarium records identify the role of long-distance spread in the spatial distribution of alien plants in New Zealand

Aim To use herbarium records to characterize important correlates of spatial spread, areal occupancy and clustering of 100 alien plant species of conservation concern in the North and South Islands of New Zealand. Location New Zealand. Methods Using herbarium data of 6294 records representing 100 alien plant species, we assessed spatial heterogeneity in the distribution of alien species by examining the role of major urban areas as sources of sampling bias. A novel method to account for spatial biases in sampling effort was developed and applied to two simple distance metrics: nearest- and furthest-neighbour spread rate. The relative importance of these two distance metrics in determining the range, areal extent and dispersion of alien species across both the North and South Islands of New Zealand was also assessed. Results The spatial distribution of herbarium records was highly clustered with a significant bias towards the more populated regions. Once sampling bias was taken into account, there was no indication that species were found closer to these urban centres than might be expected based on sampling effort. The earestneighbour spread rates were usually 1–5 km yr)1 and correlated positively with the furthest-neighbour spread rates that were an order of magnitude higher. Range and area increased and clustering decreased with higher spread rates and longer recording time span. The spread rates divided species into five groups that were clearly distinguishable in terms of the extent of their istribution and the degree of clustering. Species occurring on both islands did not exhibit similar spread rates or spatial patterns. Main conclusions The nearest- and furthest-neighbour spread rates from herbarium records can explain the area and pattern of alien plant distributions and improve the understanding of the dynamics of their spread. Five groups emerge from the spread rates in relation to a null model. Fast-spreading plants had the widest, least clustered distribution, which suggests widespread chronic problems; slow-spreading plants had localized, but dense, clustered distributions, indicating acutely problematic weeds. These patterns appear robust and may be useful in predicting the future patterns of spread and in planning long-term management strategies in New Zealand.